The Blue-Ocean Platform Map for Builders and Creators: Where to Build Now, Where to Wait, and Where to Run

• The single best “blue ocean” right now is Apple Immersive Video and spatial 3D content: Apple’s own immersive catalog is only ~47 titles (a few hours total) against an installed base of roughly 500,000–1,000,000 headsets, and Apple has explicitly pivoted to enabling third-party creators — but the economics only work if you can access a shared/rented ~$25,000 camera rig, so this is a content play with a real, AI-proof moat, not a hardware-ownership play. The second-best is the Meta Ray-Ban Display SDK (opened in developer preview May 14, 2026), the first display-bearing consumer smart-glasses product on shelves with a real third-party toolkit and near-zero apps.
• Most “AI hardware,” smart-glasses, and voice platforms are dead seas or closed gardens: Alexa Skills is the canonical cautionary tale (Amazon’s hardware unit was on pace to lose ~$10B in 2022 and it ended developer payments in mid-2023), most AI glasses have no open SDK, and ambient-recorder startups are being bought and shut by Big Tech (Meta bought Limitless in Dec 2025; Amazon bought Bee in July 2025). The genuine software gold rush is MCP servers (~100 servers Nov 2024 → 10,000 active servers and 97M monthly SDK downloads by Dec 2025 per Anthropic), where distribution is instant and hardware is irrelevant.
• Do not over-buy hardware. You need a Quest 3S ($299) and a Vision Pro (M5, ~$3,499) to validate XR; a Meta Ray-Ban Display ($799) to enter the glasses blue ocean; and — only if you commit to immersive video — access (rent, don’t buy) to a Blackmagic URSA Cine Immersive ($24,995) or a Canon EOS R7 + RF-S 3.9mm dual-fisheye (~$2,600 combined) for the low-cost VR180 tier. Everything else (smart rings, most e-ink, most glasses, automotive) is either closed, a dead sea, or adequately served by simulators.

The first-principles reframe: in 2026 the moat is no longer “code exists.” AI coding agents have collapsed the marginal cost of building an app. That means “underserved platform” is necessary but not sufficient. The durable moats have shifted to: (1) hardware/capture access (you cannot vibe-code an 8K stereoscopic camera; immersive video is defensible because the capture is expensive and scarce); (2) distribution and taste (being findable and good, not merely existing); and (3) content libraries (assets, environments, footage that take real-world time to produce). Prioritize platforms where the bottleneck is something AI cannot instantly replicate. Where the only barrier was “writing the code,” expect that barrier to be worth almost nothing by the time your product matters.

The “blue ocean = dead sea” default. The empirical first-mover literature is blunt: Lieberman & Montgomery and subsequent surveys find that order of market entry is not related to long-term survival rates; pioneer market-share “advantages” are largely a measurement artifact (tests using market share find first-mover advantage far more often than tests using profitability or survival). The academic consensus (e.g., Mitchell and colleagues) is that “the firms we discuss as having reaped first-mover advantages are often far from being the first movers; long-term market leaders are more likely to be later entrants, while failed attempts to launch what became successful markets have disappeared from memory.” Translating to app platforms: early entry only pays when the platform actually reaches scale AND users pay AND the platform owner does not cannibalize your category. Most emerging consumer-electronics platforms fail at least one of these tests.

A quantitative content-gap framework. Define the gap ratio as third-party items ÷ installed base (in millions). Mature baselines: iOS has roughly 1.8M apps across ~1.4B active devices (≈1,300 apps per million devices); Steam has 100,000+ titles across a base well over 100M monthly active users. A high ratio means saturation; a low ratio with a growing, paying base is the real opportunity. The trap is a low ratio with a tiny, non-growing, non-paying base — that is a dead sea, not a blue ocean. Apply this test to every platform below before writing a line of code.

Details by Platform Category

1. XR / Spatial Computing

Apple Vision Pro (visionOS). Installed base: IDC estimates ~390,000 units shipped in 2024 and only ~45,000 in Q4 2025; cumulative ~500,000 by end of 2025, with AppleInsider speculating possibly ~1M after the October 2025 M5 refresh (largely enterprise). Native visionOS apps crossed 1,000 within two weeks of the Feb 2024 launch (per Apple marketing SVP Greg Joswiak) and growth has been glacial since — AppleInsider called Apple’s own failure to make its stock apps native “downright neglectful” at WWDC 2025. Apple reportedly cut Vision Pro digital-ad spend by more than 95% in the US and UK during 2025 (Sensor Tower via Financial Times). The current retail model is the M5 Vision Pro at ~$3,499. Verdict on software: watch and wait — the base is too small and shrinking, and Apple cannibalizes obvious utility categories.

But Apple Immersive Video / spatial content is the standout content blue ocean. As of May 21, 2026 there are only ~47 Apple Immersive Video titles total (an independent tally by Jeff Richardson at iPhone J.D., whose 47th entry is “Real Madrid: The Weight of Greatness” — not an official Apple figure), amounting to only a few hours of runtime; individual episodes run ~8 minutes (e.g., “Elevated: Switzerland,” 9to5Mac, Feb 2026). Apple’s flagship series are threadbare: “Wild Life” and “Adventure” had only 4 and 5 episodes and “Elevated” had 1 at the time of reporting (Bloomberg’s Mark Gurman, Aug 2025). Gurman reports Apple is deliberately slowing releases because the content is “exceptionally expensive” and the install base is under 1M. Crucially, Apple has pivoted to enabling third parties: it previewed 7 new third-party immersive films (Sept 2025) shot on Blackmagic gear, and ran a two-day immersive-production seminar in Cupertino (CineD, 2026) explicitly addressing creator monetization. This is a textbook “pipeline just opened, near-zero supply, platform owner actively recruiting you” moment — the defensible kind, because the capture hardware is a real barrier.

The capture hardware reality: the Blackmagic URSA Cine Immersive launched at $29,995 and is now $24,995 (with a $26,495 100G live variant and a ~$1,495–$1,645 live encoder); it is the only camera purpose-built for the Apple Immersive format, paired with DaVinci Resolve Studio for post. The affordable tier is the Canon EOS VR system: the RF-S 3.9mm F3.5 dual-fisheye lens ($1,099) on a Canon EOS R7 (~$1,500 body), roughly $2,600 combined, produces VR180 3D — lower quality than Apple Immersive but usable for Quest/DeoVR. (Canon’s higher-end RF 5.2mm F2.8 L dual-fisheye runs ~$1,999 and pairs with full-frame bodies like the R5.) There are already 52,100+ VR videos on DeoVR, but developer-creator Anthony Maës notes their quality “is not on par with Apple’s offerings” and most are absent from Apple’s store — the gap is premium, Apple-format immersive content, not raw quantity.

Meta Quest (Horizon OS). This is the only VR platform with a real consumer base and real money — but it is a mature, stagnant market, not a blue ocean. Active installed base is estimated at ~25–35M units across Quest 2/3/3S. Store revenue: Meta cited “over $2 billion” cumulatively in both Sept 2023 and again more than a year later (Road to VR), i.e., the $3B milestone had not been crossed — a stagnant marketplace. At GDC 2026, Meta games director Chris Pruett disclosed that 2025 Horizon Store gross revenue was up “very slightly” YoY (roughly matching ~1% flat industry growth); over 100 titles grossed $1M+ in 2025; and Horizon+ passed 1M subscribers with a 100+ game catalog and paid ~$20M to participating developers. Pruett was candid to Lowpass: “This is the roughest period I have ever seen… revenue in the Quest store was still up slightly in 2025,” while noting Meta “shut down several of our studios.” Total app count is estimated at 10,000+ (Treeview, analyst aggregation). Commission is 30%; the Meta Developer Program is free. Meta also cut ~10% of Reality Labs XR staff in early 2026 and paused its third-party Horizon OS headset program (Dec 2025) — bearish signals. Verdict: build only if you have genuine VR game-design chops; it is competitive and flat, not untapped. The real bottleneck is 3D art and comfort tuning, which AI does not yet solve well — a moat, but a labor-intensive one.

Android XR / Samsung Galaxy XR (Project Moohan). Launched Oct 21, 2025 at $1,799 (controllers +$250); built on Android XR by Samsung/Google/Qualcomm, using the Jetpack XR SDK (Kotlin/Compose), OpenXR, Unity, and WebXR. Every Android app runs out of the box, so the “app gap” is partly auto-filled — which cuts both ways: low differentiation for 2D ports (they render as small floating windows). Google shipped Android XR SDK Developer Preview 3 (Dec 2025) with a hardware emulator to seed developers ahead of consumer hardware. Verdict: watch closely, prototype now on the free emulator (no device needed for early work). The open toolchain plus Google/Samsung distribution makes this the most credible medium-term XR bet, but there is essentially no installed base yet. Build spatial-native experiences, not 2D ports.

Valve Steam Frame. Announced Nov 12, 2025; standalone SteamOS headset (Snapdragon 8 Gen 3, dual 2160×2160, eye tracking) with a 6GHz wireless PC-VR dongle. As of late July 2026, no official price (analyst range $899–$1,199, pushed up by the DRAM shortage that put the companion Steam Machine at $1,049) and no ship date beyond “summer 2026.” The “Great on Frame” verified section is live and filling. Verdict: watch — Steam’s catalog auto-populates via FEX/Proton, and Valve’s audience pays, but it is unreleased. Position by building/porting PCVR content that will earn a “Great on Frame” badge at launch.

Others. Bigscreen Beyond (tiny, tethered enthusiast), Play For Dream (premium Android-XR-adjacent), Varjo (enterprise), Pico (strong in China, weak in the West) — all niche or regional; none is a general blue ocean for an indie.

2. Smart Glasses / AI Wearables — the SDK-existence test is decisive

The category shipped ~7.25M units in 2025 (~50% of all XR hardware), with Meta holding 70%+ of AI-glasses volume. But most have no open third-party SDK, which means the blue ocean is literally un-enterable until it opens. This is the “SDK doesn’t exist yet” trap: you cannot be early on a platform that has no door.

• Meta Ray-Ban Display ($799, on sale since Sept 2025) — THE new blue ocean. On May 14, 2026, Meta opened the display to third parties in developer preview via the Wearables Device Access Toolkit (native Swift/Kotlin SDK) plus a new Web Apps path (HTML/CSS/JS), both exposing the Neural Band sEMG wristband for input. This is the first consumer display-glasses product on shelves with a real third-party display SDK and effectively zero third-party apps. During preview, developers can share builds with up to 100 testers; public publishing opens later in 2026. Meta has not yet specified app-review rules or a monetization/store model — that is the central risk. Verdict: build now, cheaply, via the Web Apps path; the barrier to entry is low (web skills) and being early here is genuinely possible.
• Ray-Ban / Oakley Meta (non-display). The Device Access Toolkit lets you pipe camera/audio/voice to a companion phone app; useful but constrained.
• Snap Spectacles / Lens Studio. The most mature AR-creator ecosystem: per Snap’s official Lens Fest 2025 newsroom post, 400,000 people have built more than 4 million Lenses, and Lens Games now reach 175 million people monthly (up 130% YoY). Payouts exist (Lens Creator Rewards / Top Performer, up to $7,200/month for a top Lens; Lens+ Payouts added Oct 16, 2025), but cumulative “Top Performer” payouts are only “more than $1M” — real but small. Camera Kit reaches 68M+ MAU and dropped mandatory branding in 2025. Verdict: strong for content-style AR creators, weak for those wanting large per-project revenue.
• Xreal (One/One Pro, Air, Eye). Open developer program (10,000+ members), Unity-based XREAL SDK under the Unity XR umbrella; 350,000+ glasses sold. Display-in-air, not full AR. The Aura (Android XR) glasses launch Fall 2026 under ~$1,500. Verdict: viable niche for productivity/media apps.
• Brilliant Labs Frame/Halo. Fully open-source, the most hackable — best for research/experimentation, tiny user base.
• Even Realities G1/G2, Rokid (~$350), Halliday. Mostly closed or minimal SDK; discreet-display niche.
• Amazon glasses / OpenAI-io device. Rumored/unreleased; no SDK — cannot enter yet.

3. AI Hardware / Voice-First / the LLM-adjacent layer

• MCP (Model Context Protocol) — the real software gold rush, hardware-free. From ~100 servers at launch (Nov 2024) to 5,800+ by April 2025 to, per Anthropic’s official MCP blog (Dec 9, 2025), “over 97 million monthly SDK downloads, 10,000 active servers and first-class client support across major AI platforms like ChatGPT, Claude, Cursor, Gemini, Microsoft Copilot, Visual Studio Code.” MCP was donated to the Linux Foundation’s Agentic AI Foundation in Dec 2025. Caveat: registries list 90,000+ catalogued servers but a 2025 measurement study validated only ~8,060 active out of 17,630 collected, and 50%+ are inactive/low-value — a crowded field, but distribution is instant and the tooling barrier is trivial. Verdict: build now — this is the lowest-friction, highest-reach opportunity in the report, and AI-cost-collapse helps you here (you can ship many servers fast). The moat is picking a genuinely useful, hard-to-replicate integration, not the coding.
• Ambient recorders (Limitless, Plaud, Bee, Omi). Consolidating fast: Meta bought Limitless (Dec 2025, ended pendant sales), Amazon bought Bee (July 2025), Rewind shut desktop capture (Dec 2025). Plaud is the last big independent — $250M ARR, 1.5M+ users (April 2026), targeting $500M in 2026 (Bloomberg, June 2026) — but has no open third-party app platform. Omi is open-source ($89) with a community plugin ecosystem — the only real entry point for indie builders. Verdict: the platform layer is largely closed; build Omi plugins if anything.
• Alexa Skills — the cautionary tale, quantified. Amazon ran developer reward programs, then wound them down: it ended AWS-credit rewards and monthly cash payments to skill developers as of June 30, 2023. Business Insider (Nov 2022), interviewing “a dozen current and former employees on the company’s hardware team,” reported the hardware unit was on pace to lose $10B in 2022; a former employee called Alexa “a colossal failure of imagination” and “a wasted opportunity.” Near-zero monetization despite a massive installed base. This is the archetype of a large “ecosystem” that never paid developers — the exact fate to fear on AI glasses and voice devices.

4. Automotive / In-Car Apps

Android Automotive OS (AAOS, “Google built-in”) is on 50+ vehicle models with 300+ apps on the car Play Store (Google I/O 2025); Android Auto reaches ~250M compatible vehicles. Google’s “Car ready mobile apps” program (Feb 2025) auto-distributes eligible video/game/browser apps to AAOS cars with little extra work. Verdict: genuine app gap, but high friction — access to vehicle-data APIs (VHAL) and OEM relationships are gated, and safety review is strict. Best for parked-experience apps (video, games, browsers) via the Car-ready program. Tesla remains a closed in-car app situation; Rivian/Lucid/Polestar/Volvo stores are largely closed or OEM-curated. Watch and wait unless you target the parked-app category via Google’s program.

5. Handheld Gaming and Console-Adjacent

• Playdate. 80,000+ units sold (Nov 2025) at $229. The Catalog store paid out just over $1M to developers across 289,000+ game sales (April 2025), average game price ~$3.60, 262 Catalog titles plus 800+ on itch.io. The SDK is free and beloved; the crank is a unique creative constraint. Verdict: build for love and craft, not money — the payouts are tiny (a charming dead sea economically), but it is a superb portfolio platform with low competition and a passionate community. A simulator ships with the SDK, but the crank/1-bit screen demand a real unit (~$229) to ship well.
• SteamOS/Steam Deck & Steam Frame. Deck-verified niche utilities are underserved; see Steam Frame above.
• Analogue / retro-handhelds. Hardware-restoration niche, not a software platform.

6. Smart TV / Streaming OS and Ambient/Art Content

The Samsung Frame Art Store is a real, growing content channel (partners include The Met, MoMA, Tate, and Etsy artists; described as a “multi-billion-dollar business” with millions of Frame units, 1,400+ works catalogued years ago and far more now). But it is curated/closed — you sell through partnerships, not open self-publish. TV OS app stores (Tizen, webOS, Roku, Google TV, Fire TV) are underserved for quality apps but painful to build for and low-paying. Ambient/screensaver/art content is a genuinely underserved category across Frame, Apple TV screensavers, and Vision Pro Environments — and it plays directly to a content creator’s strengths. Verdict: pursue ambient/art content, but expect to negotiate into curated stores rather than self-publish.

7. Wearables Beyond Phones

• WearOS watch faces. Samsung closed global WearOS watch-face submissions on the Galaxy Store (only China supported), per developer-forum complaints — a friction/closure signal. Google’s Watch Face Format is the path; underserved but low-paying.
• Apple Watch complications. Mature, Apple-dominated.
• Smart rings (Oura, Ultrahuman, Circular). Largely closed — no meaningful third-party on-device app platform; Oura exposes data via API/partners, not installable apps. Verdict: not enterable as an app platform.
• Hearables/earbuds SDKs. Mostly closed or trivial.

8. E-ink / “Calm Tech”

• reMarkable Paper Pro ($629) now ships an official developer SDK (a Yocto/x86_64 cross-compile toolchain) plus a Developer Mode for shell access — a real opening, though the default OS refuses third-party apps and the audience is small and single-purpose. Verdict: early-mover possible for niche utilities, tiny market.
• Boox (Note Air 4C ~$500) is Android-13-based with the Play Store — fully open; build standard Android apps tuned for e-ink.
• Daylight DC-1, TRMNL, Kindle Scribe. TRMNL and e-ink dashboards are a charming maker niche; Kindle Scribe is closed. Verdict: calm-tech is a real aesthetic trend but a small-money niche.

9. Robotics / Embodied

LeRobot (Hugging Face) open-source robotics, Unitree, robot vacuums (mostly closed SDKs), drones. Verdict: high learning curve, mostly research-stage or closed; not a near-term monetizable app blue ocean for a generalist, but a strong “watch” for the AI-agent-meets-hardware wave.

10. Creative/Pro Tool Plugin Ecosystems (underrated, hardware-free)

Blender add-ons, DaVinci Resolve, Figma plugins/widgets, Raycast extensions, Obsidian plugins, Home Assistant integrations, and Bambu Lab’s MakerWorld (3D-model sharing with a points/rewards program) are all open, low-friction, and monetizable — and they play to taste and craft, which AI-cost-collapse does not erase. Verdict: build now — these are the most reliable “small but real money, low barrier” opportunities, especially Figma, Raycast, Obsidian, and MakerWorld.

11. Brand-new 2025–2026 launches to be an early mover on

The three freshest openings with open (or newly opened) SDKs and near-zero third-party supply are: the Meta Ray-Ban Display toolkit (May 2026), the reMarkable Paper Pro developer SDK (2025), and Android XR (Developer Preview 3, Dec 2025, hardware barely shipping). MCP is the fresh software standard (late 2024 → mainstream 2025–2026). These are where “early” is still literally possible.

Recommendations (staged, with thresholds)

Stage 0 — build immediately, near-zero hardware risk (this quarter):

1. MCP servers for genuinely useful, hard-to-replicate integrations. No hardware. Leading indicator to watch: broadening of MCP write/action support in ChatGPT/Claude as the monetization trigger.
2. Meta Ray-Ban Display Web Apps ($799 device to test properly; start in the preview). Aim to be among the first 100 apps. Threshold to double down: Meta announcing a store + monetization model.
3. Creative-tool plugins (Figma, Raycast, Obsidian, DaVinci Resolve, Blender, MakerWorld). No special hardware.

Stage 1 — content creation with a real moat (this half-year): 4. VR180/immersive video starting on the Canon EOS R7 + RF-S 3.9mm (~$2,600) to learn the craft cheaply; graduate to renting a Blackmagic URSA Cine Immersive ($24,995 to buy — do not buy until you have a paying pipeline) to produce Apple-format immersive content. Threshold to buy the Blackmagic: a signed distribution/sponsorship deal or a repeatable per-title revenue path via Apple’s monetization program. 5. Ambient/art/environment content for Samsung Frame, Apple TV, and Vision Pro Environments.

Stage 2 — watch-and-wait; prototype on emulators, buy hardware only at the trigger: 6. Android XR (prototype on the free emulator now; buy Galaxy XR at $1,799 only if a second, cheaper OEM device ships or the installed base crosses ~1M). 7. Steam Frame (build/port PCVR for a launch “Great on Frame” badge; buy at launch if priced under $1,000).

Avoid (dead seas / closed): Alexa Skills; smart-ring apps; standalone ambient-recorder platforms (build Omi plugins only); Vision Pro utility apps (Apple cannibalizes); most AI glasses without SDKs; WearOS watch faces as a business.

Hardware Purchase Plan (unlimited budget, strong bias against waste)

• Buy now: Meta Quest 3S ($299) — cheapest way to validate VR comfort, passthrough, and motion sickness (the things a simulator cannot validate); Meta Ray-Ban Display ($799) — required to test the new display SDK’s real-world legibility and Neural Band input; Apple Vision Pro M5 (~$3,499) — required to judge immersive-video quality and eye/hand interaction, which no simulator reproduces.
• Buy only when committed to a content line: Canon EOS R7 + RF-S 3.9mm (~$2,600) for VR180; access to Blackmagic URSA Cine Immersive ($24,995) via rental first.
• Do NOT buy: Samsung Galaxy XR ($1,799, use the emulator until the base grows), any smart ring (closed), most AI glasses without SDKs, reMarkable/Boox unless you commit to an e-ink product.
• Developer-account costs & gates: Apple Developer Program $99/yr (visionOS submissions reviewed); Meta Developer Program free (Horizon Store is curated/reviewed, 30% cut); Google Play $25 one-time (AAOS/Android XR; AAOS adds strict safety review); Snap Lens Studio free; Playdate SDK free (Catalog takes 25%); MCP/plugin ecosystems free.
• What only physical hardware can validate: comfort, weight, latency, motion sickness, real-world lighting/passthrough quality, display legibility outdoors, and battery life. What simulators/emulators handle fine: UI layout, app logic, most Android XR and AAOS early development, MCP and plugin work (entirely hardware-free). Cloud device farms (e.g., AWS Device Farm, Firebase Test Lab) cover standard Android/AAOS but do not meaningfully cover XR headsets or glasses — physical units are unavoidable for those.

Specific Project Ideas (prioritized by expected value)

1. MCP server bridging a high-value, poorly-served data source to Claude/ChatGPT — e.g., “personal-finance-across-institutions” or “local-government-open-data” with real write actions. Missing because most of the 10,000 active servers are trivial read-only wrappers. Low build cost; monetize via a hosted SaaS tier. Highest EV.
2. Meta Ray-Ban Display Web App: live teleprompter + “conversation-notes HUD” showing talking points and name reminders in-lens, driven by the phone. Missing because the display SDK is three months old. Web skills only; monetize via subscription once Meta’s store opens.
3. Apple Immersive Video niche series Apple isn’t making — e.g., “immersive trades apprenticeship” (welding, watchmaking, procedure-POV), or “immersive live music from small venues.” Evidence: Apple’s own catalog is ~47 titles and it is actively recruiting third parties. Rent the Blackmagic; monetize via Apple’s creator program / licensing.
4. VR180 “comfort travel” and “family memory” packages shot on the Canon R7 rig, sold on DeoVR/Quest — the existing 52,100 DeoVR videos are largely low-quality; premium, tasteful VR180 is scarce.
5. Figma/Raycast/Obsidian plugin wrapping an AI agent for a specific workflow (e.g., a Raycast extension that runs local MCP tools). Proven monetization, low barrier.
6. MakerWorld parametric 3D-model line (e.g., modular desk organizers, e-ink dashboard enclosures for TRMNL/Boox) — earns MakerWorld points and drives print sales.
7. Ambient art collections for Samsung Frame / Apple TV / Vision Pro Environments — original generative or photographic “living wallpaper” sets; underserved and taste-driven.
8. Omi (open-source pendant) plugins — e.g., a real-time “meeting-action-item extractor” — the only open ambient-AI entry point after the Limitless/Bee acquisitions.
9. reMarkable Paper Pro utility via the new SDK — e.g., a distraction-free RSS/read-later or habit tracker — early-mover on a just-opened SDK.
10. Playdate crank-native micro-game — build for craft, portfolio, and community, not the ~$3.60 average revenue.
11. Android XR spatial-native app (built on the emulator now) for a use case 2D ports can’t do — spatial data-visualization or a room-scale training tool — ready to ship the day the base grows.
12. “Great on Frame” launch-day PCVR port of an existing quality experience, to capture the badge and the early Steam Frame audience.

Caveats

• Vision Pro installed-base and immersive-title counts are estimates, not official disclosures: IDC’s ~390k (2024) / ~45k (Q4 2025) shipments and the ~47-title tally (iPhone J.D., which includes not-yet-released titles) are the best available but unconfirmed by Apple. Apple’s “150+” launch figure conflated ordinary 3D movies with true immersive video and should not be read as the immersive catalog.
• Meta’s store app count (10,000+) is an analyst aggregation; Meta stopped publishing precise counts, and a separate commercial dashboard lists only ~5,900 titles. Meta’s revenue disclosures ($2B+ cumulative; ~$20M Horizon+ payouts; 100+ $1M titles) are official but selectively framed.
• The Ray-Ban Display SDK lacks a published monetization/store model and app-review rules — the biggest risk to that thesis; being early is only valuable if Meta opens a paying store. Position, prototype cheaply, but don’t bet a year on it before the store exists.
• First-mover advantage is empirically weak — treat every “build now” as a bet on the platform reaching scale, and keep projects cheap (leaning on AI coding) so any single failure is survivable.
• Steam Frame pricing/date and several 2026 hardware prices are moving targets amid a severe DRAM shortage; verify prices at purchase time.
• Where reliable data simply does not exist (e.g., total hours of Apple Immersive Video, precise VR180–3D-only counts, exact current Quest store size), I have flagged the gap rather than guessed.