Microsoft Unveils 7 AI Models Beating Claude as Coinbase Backs GENIUS Act ETF

Crypto News

Microsoft used the opening day of its annual Build developer conference to introduce seven proprietary artificial intelligence models, positioning itself as a frontier AI builder rather than purely an OpenAI backer. Microsoft AI chief executive Mustafa Suleyman described the family as a new era in AI designed to keep developers in control and on the frontier. The lineup spans reasoning, coding, image generation, transcription, and voice synthesis. Each system was reportedly trained from a clean data lineage and engineered for efficiency. The release marks the company's most ambitious independent modeling push yet, signaling intent to compete directly with Anthropic and Google across the full enterprise stack.

At the center of the rollout sits MAI-Thinking-1, the company's flagship text foundation model focused on reasoning. Independent blind evaluations reportedly favored MAI-Thinking-1 over Anthropic's Claude Sonnet 4.6, while benchmark scores placed it level with Claude Opus 4.6 on the SWE Bench Pro coding test. Microsoft also reported a 97 percent result on AIME 2025, a measure of advanced problem-solving ability. Suleyman framed the model as proof that Microsoft can ship competitive frontier systems built in-house, drawing comparisons to blockchain infrastructure races where vertical integration ultimately defined market leadership across the previous technology cycle.

Alongside the reasoning model, Microsoft introduced MAI-Code-1-Flash, a lightweight coding model tuned for GitHub Copilot and Visual Studio Code, and MAI-Image-2.5 with a Flash variant that the company says outperforms Google's Nano Banana 2 on image-editing leaderboards. The lineup also includes MAI-Transcribe-1.5, supporting forty-three languages, and MAI-Voice-2, a speech-generation system capable of producing natural-sounding output across fifteen languages from a brief audio sample. Together the suite covers most modalities developers need in production, suggesting Microsoft intends to offer a full-stack alternative across enterprise software, developer tooling, and consumer creative applications worldwide.

Pivoting to digital asset markets, Coinbase announced an undisclosed investment in ProShares' GENIUS Money Market ETF, ticker IQMM, a fund designed to hold the kinds of assets that qualify as reserves for payment stablecoins. The vehicle invests exclusively in short-term US Treasury securities and cash-equivalent instruments with maturities of ninety-three days or less. As one of the primary infrastructure providers for Circle's USDC, the exchange has direct interest in expanding the pool of regulated, liquid instruments that can manage stablecoin reserves at scale. The commitment signals confidence that compliant reserve-management products will see accelerating institutional demand throughout 2026.

The GENIUS Act, enacted in June 2025, formalized the rules governing US dollar-pegged tokens, requiring issuers to back their tokens with highly liquid assets including cash, bank deposits, and short-term Treasury securities. IQMM was launched in February as one of the first exchange-traded funds tailored specifically for stablecoin reserve management, providing publicly traded exposure to qualifying instruments. The Coinbase commitment aligns the exchange's growing stablecoin business and treasury operations with that statutory blueprint. It also illustrates how regulated funds are emerging as a preferred venue for issuers seeking compliant yield on idle balances within the broader DeFi and payments ecosystem.

Attention is now shifting to the Digital Asset Market Clarity, or CLARITY, Act, which would define federal regulator roles and establish market structure rules for digital assets including Bitcoin and major altcoin tokens. The bill advanced through the Senate Banking Committee last month and is positioned for a full floor vote, though progress has been uneven. Recent additions on stablecoin yield provisions have sharpened debate over whether issuers should pay interest to holders. White House crypto adviser Patrick Witt previously targeted the period around the July 4 holiday for advancing market-structure legislation, but ongoing disputes over ethics and conflict-of-interest provisions leave that timeline uncertain.

A common thread connects both narratives: large incumbents are racing to control the foundational layers of the next technology cycle. Microsoft is internalizing the model stack that powers its software empire, while Coinbase is buying into the regulated plumbing that will channel stablecoin reserves under the GENIUS Act. Whether the catalyst is artificial intelligence sovereignty or US crypto market structure, the dominant narrative this quarter is institutional consolidation around proprietary infrastructure. Capital and policy are converging on the same outcome — that the entities owning the rails, the reserves, and the models will define how the next phase of digital finance and computing unfolds.