Alphabet (GOOGL) on Solana

Alphabet has expanded its AI chip partnership with Marvell Technology, with the agreement including a warrant that would allow Google to acquire a significant ownership stake in Marvell valued at up to $12.18 billion. The option is tied to directed chip design and manufacturing business Google sends Marvell's way, giving Alphabet both supply-chain leverage and potential financial upside as the relationship scales.

The move signals a deliberate effort to reduce dependence on dominant GPU suppliers by cultivating an external partner alongside its own in-house TPU development. The partnership is aimed at expanding compute capacity for Google Cloud and Gemini services, though the capital-intensive arrangement has drawn analyst attention to Alphabet's record capex trajectory and the margin pressure that could follow if AI infrastructure spending outpaces revenue growth from cloud and AI services.

Alphabet is deploying Gemini 3.7 Flash, a model focused on coding automation and workflow efficiency, across Google Cloud, Google Workspace, and developer tools. The rollout reflects a deliberate strategy to broaden AI adoption by embedding capable—but not premium-priced—models into products already used by billions of people, rather than relying solely on high-cost, flagship offerings to drive revenue.

The release comes alongside a delayed rollout of the higher-end Gemini 3.5 Pro, a gap that has drawn scrutiny from investors watching competitive pressure from open-source models and lower-cost AI alternatives from Chinese developers. Key indicators analysts are tracking include paid Google AI tier uptake, Cloud and Workspace AI feature engagement, and AI-influenced deal flow through the rest of 2026.

Berkshire Hathaway sharply increased its Alphabet stake in the second quarter, raising its position by 83% to roughly 106 million shares worth $37.9 billion at the end of June, vaulting Google's parent company past most of the conglomerate's portfolio to become its third-largest U.S.-listed equity holding behind only Apple and American Express. The move is partly attributable to a $10 billion private stock purchase announced in early June, in which Alphabet sought fresh outside capital to help finance its aggressive AI infrastructure buildout — a bet Berkshire chose to take on at scale rather than in the open market.

The timing carries symbolic weight. Berkshire's endorsement as a quintessentially value-oriented institution — historically skeptical of high-capex technology plays — lands at a moment when markets have punished Alphabet's stock for its towering capital expenditure commitments. That Berkshire, now deploying cash under CEO Greg Abel, views Alphabet's AI infrastructure spending as warranting a near-$38 billion conviction signals a meaningful vote of confidence in the company's long-term competitive position in AI, cloud, and advertising.

Alphabet launched the Pixel 11, Pixel 11 Pro, Pixel 11 Pro XL, and Pixel 11 Pro Fold on August 12, with core models set to go on sale August 20. The defining feature of the lineup is Gemini Intelligence, an on-device AI system that operates across the Android operating system rather than as a standalone chatbot. It reads user activity to connect messages, calendar, and maps, suggests contextual actions, initiates restaurant bookings, and includes scam detection — working proactively across more than 40 apps without requiring a user prompt.

The launch arrives just weeks before Apple is expected to ship a rebuilt Siri that runs on Gemini, placing Alphabet in the unusual position of competing directly with one of its largest AI customers. Devices chief Rick Osterloh acknowledged pricing pressure from a broader memory shortage — RAM costs have surged from $2.80 to $12 per gigabyte — though no final Pixel 11 pricing was disclosed at launch. Counterpoint Research notes that AI features currently rank sixth or seventh among consumers' stated reasons to upgrade a phone, though the ranking is climbing as agentic capabilities become more visible.

Alphabet has elevated Koray Kavukcuoglu to run Google DeepMind's day-to-day AI research and operations from Mountain View, consolidating power at its California headquarters as the company tries to close ground on OpenAI and Anthropic. DeepMind co-founder Demis Hassabis, who built the lab into one of the world's premier AI research institutions, steps back from operational leadership to become chairman of Google DeepMind and Alphabet's Chief Scientist, remaining based in London. The restructuring is designed to untangle friction left by the 2023 merger of Google Brain and DeepMind, two organizations that operated across different continents with different cultures and frustrated researchers on both sides.

Kavukcuoglu inherits a difficult competitive position: Alphabet has pledged up to $205 billion in 2026 capital expenditures to support AI ambitions, yet Google's flagship Gemini model is reportedly months behind schedule on its most powerful iteration, and the company is widely viewed as secondary to rivals in AI coding automation — the first clear commercial application to define the current era. The transition comes amid a sustained talent drain, with prominent departures including legendary engineer Jeff Dean and researcher David Silver, who secured $1.1 billion in seed funding for his new startup Ineffable Intelligence. As the only remaining Gemini co-lead, Kavukcuoglu faces immediate pressure to stabilize the team and accelerate model delivery against OpenAI's and Anthropic's momentum.

Alphabet is facing a significant reshaping of its AI leadership as some of the researchers who defined the field depart for frontier labs and new ventures. Jeff Dean, Google's chief scientist and a 27-year company veteran who co-invented foundational infrastructure like MapReduce and TensorFlow, is leaving to launch Discovery Loop — a startup backed by Google and joined by fellow Google luminaries Sanjay Ghemawat, Oriol Vinyals, and Quoc Le. Separately, Demis Hassabis is stepping down as DeepMind CEO to become chairman, shifting into a long-term AGI-focused chief scientist role at the Alphabet level. The departures follow Noam Shazeer — co-author of the landmark 2017 "Attention Is All You Need" paper — leaving for OpenAI after Alphabet reportedly spent roughly $3 billion to rehire him, and Nobel laureate John Jumper departing DeepMind for Anthropic.

The exodus reflects a structural tension between Alphabet's commercially successful AI scale-up and the priorities of researchers drawn to cutting-edge model development. Google Cloud posted 82% revenue growth in Q2 2026 and Gemini is now used by 90% of Fortune 100 companies, strengthening the commercial case for infrastructure and enterprise AI. But analysts note that top researchers increasingly view competitors like Anthropic and OpenAI — companies that Google's own TPU sales help power — as the places where frontier work happens. The result is Alphabet simultaneously expanding its AI empire in revenue terms while ceding some of the human capital that made that empire possible.

Washington D.C. has emerged as a key front in a regulatory fight that could constrain Waymo's driverless ride-hailing expansion. Uber is backing proposed DC legislation that would impose a $1 million application fee and a $0.15-per-mile surcharge on autonomous vehicle operators — a structure that directly targets Waymo's driverless-only service model rather than Uber's own hybrid approach using human drivers alongside AV partners. In New Jersey, Uber lobbyists have also circulated draft legislation that would require human drivers to account for at least 85% of all rideshare trips on any platform offering robotaxi services over the next three years, a threshold that would effectively cap Waymo's fleet share in that market.

The labor dimension is central to the Washington fight. ATU Local 689, the transit union representing DC-area workers, has warned that unconstrained AV expansion could eliminate hundreds to thousands of local jobs and is pressing for phased deployment rules and worker transition protections. For Alphabet, which has invested heavily in Waymo's fully autonomous model and has been scaling commercial robotaxi operations in multiple cities, the DC and New Jersey regulatory templates represent a meaningful risk: if adopted broadly, per-mile tax structures and human-driver minimums would raise Waymo's operating costs and legally limit its market share at precisely the moment it is trying to demonstrate commercial viability at scale.

CNBC's Jim Cramer said he was "shocked" by Wall Street's reaction to Alphabet's Q2 2026 earnings, specifically the stock's 7% plunge on July 23rd — wiping out roughly $255 billion in market cap — after the company raised its 2026 capital expenditure guidance from $180-190 billion to $195-205 billion. Cramer summarized the market mood bluntly: "if you spend, your stock is in the penalty box," pointing to investor frustration that massive AI infrastructure outlays haven't yet produced a clear return on investment. The selloff came despite Alphabet reporting revenue of $119.80 billion — well ahead of estimates — though adjusted EPS of $2.85 missed the $2.89 consensus, and free cash flow turned negative for the first time since the company's 2004 IPO.

While Cramer acknowledged standout performance in Google Cloud, which posted 82% revenue growth to $24.8 billion in the quarter, he indicated he would look elsewhere in tech until Alphabet can demonstrate concrete AI ROI to justify its capital intensity. The reaction highlights a broader investor tension around hyperscaler AI spending: strong top-line results are being discounted when capex guidance signals years of heavy investment ahead with uncertain payback timelines.

Google DeepMind has disbanded the dedicated team behind AlphaFold, the landmark protein-structure prediction system, reassigning researchers to work on Gemini, Alphabet's large language model platform. The protein-folding research itself is not being abandoned; it will continue as part of broader scientific programs spanning enzyme design, genomics, drug discovery, and nuclear fusion, but without a standalone unit driving it.

The restructuring reflects Alphabet's strategic judgment that a unified, large-scale model platform delivers more business leverage than specialized research silos. By concentrating DeepMind talent on Gemini, the company aims to monetize AI capabilities through Google Cloud and compete more directly with Microsoft, OpenAI, and Meta in supplying general-purpose AI systems to enterprise and scientific customers. The move underscores a recurring tension at Alphabet between maintaining frontier research prestige and deploying AI at commercial scale.

Alphabet's advertising and cloud segments are driving renewed optimism about the stock's long-run trajectory, according to an analysis published by Motley Fool via Yahoo Finance. Google and YouTube combined generated $81.6 billion in advertising revenue in Q2 2026, up 14.4% year-over-year, with AI Overviews and AI Mode credited for lifting engagement and advertiser demand. Google Cloud posted $24.8 billion in revenue for the same quarter — 82% above the prior year — while operating income for the segment surged 212% to $8.8 billion. The cloud backlog now stands at $514 billion, providing multi-year revenue visibility.

The analysis also cites Alphabet's Gemini ecosystem as a compounding growth driver, with 950 million monthly active users and adoption across roughly 90% of Fortune 100 companies. Waymo's robotaxi expansion into additional markets rounds out the broader portfolio case. Against a five-year return of approximately 145%, the piece argues these business lines give Alphabet a credible path to doubling market value by the end of the decade, though the author disclosed holding positions in the stock.