Tech stocks today: Anthropic releases its newest model, Claude Opus 4.7

Anthropic launched Claude Opus 4.7, highlighting gains in advanced software engineering and long-running coding tasks, but said it is still less capable than its Mythos model. Taiwan Semi reported a 58% jump in first-quarter profit, reinforcing that AI demand remains robust. The article also underscores intensifying OpenAI-Anthropic competition, with OpenAI’s revenue chief alleging Anthropic’s $30 billion run rate is overstated by roughly $8 billion.

Analysis

The near-term winner is not “AI” broadly but the infrastructure layer that monetizes agentic coding workloads. TSM’s print is the clearest confirmation that inference-heavy and training-heavy demand are still running ahead of supply discipline, which should keep pricing power intact for the foundry, advanced packaging, and HBM ecosystem over the next 2-3 quarters. Second-order, the more frontier models are used for long-running coding tasks, the more demand shifts from one-off API calls to persistent compute consumption, favoring suppliers with capacity visibility over application-layer names whose differentiation can be rapidly competed away.

OpenAI’s attack on Anthropic’s reported growth is strategically important even if the accounting critique is partially self-serving. It signals the market is moving from “who has the best model” to “who can convert usage into defensible enterprise revenue,” and that usually compresses multiples for the smaller pure-play unless it can prove retention and margin. The implied pressure is on cloud partners too: if customer concentration and rev-share economics are being scrutinized, investors may start demanding cleaner evidence of net revenue quality, which is mildly positive for MSFT and AMZN only if their attached AI distribution remains sticky and not merely a pass-through to model vendors.

The contrarian read is that the market may be over-attributing any weakness in TSLA to geopolitics and underpricing the durability of AI capex. If AI demand stays robust, TSLA’s AI-chip tease is a noise event unless it translates into manufacturing execution or margin leverage; otherwise the stock remains hostage to sentiment and positioning rather than fundamentals. The bigger risk to the AI tape is not demand collapse, but a credibility reset in private-market revenue narratives, which could trigger a 1-2 month de-rating across late-stage AI beneficiaries before the hard earnings data reasserts itself.

Key Decisions for Investors

• Long TSM vs short an AI software basket for the next 1-2 quarters: TSM has the cleanest evidence of real monetization, while app-layer names remain vulnerable to multiple compression if growth quality gets questioned.
• Add to MSFT on 4-8 week weakness only if enterprise AI trust messaging continues to gain share; upside is modest but downside is buffered by distribution and balance sheet strength.
• Buy AMZN on dips as a relative beneficiary of enterprise AI spend normalization and cloud attach; use a 3-6 month horizon, targeting a rerating if investors focus on net rather than gross AI economics.
• Avoid chasing TSLA on AI-chip headlines; prefer a tactical short or put spread if momentum fades, since the stock is still trading more on narrative than on verifiable AI cash flow.
• If Anthropic-related skepticism spreads to private AI peers, use it to add selective hedges against the broader AI thematic basket for 30-60 days, as the market may temporarily punish duration and revenue-quality risk.