Skip to main content Skip to header search Skip to header search

Message important : Attention aux escroqueries par usurpation d'identité

Des messages et des publicités frauduleuses circulent actuellement sur les plateformes de messagerie, y compris WhatsApp, en usurpant l'identité de la marque Candriam et de ses employés. Veuillez noter que Candriam n'offre jamais de recommandations d'investissement ou de conseils financiers par le biais de canaux tels que les médias sociaux ou les plateformes de messagerie.

Technology: conviction beyond the correction

Following strong performance in recent years, the information technology sector and the broader AI trade are experiencing significant volatility. Investors are increasingly weighing the risks surrounding the next phase of AI development. Yet demand remains strong, earnings expectations continue to rise and AI adoption is broadening.

Artificial intelligence is entering a more demanding phase. The scale of infrastructure spending has raised expectations, and investors increasingly want evidence that these commitments will generate compelling returns. This scrutiny can create opportunities when uncertainty overshadows progress in the underlying businesses.

 

Recognising the risks

In our view, three main risks are currently defining investor sentiment. The first concerns return on invested capital. Rapid technological change, shorter hardware lifecycles and competition from open-source models make the payback on AI infrastructure harder to predict. Growing usage must translate into sufficient profits to justify the investment.

The second relates to financing and free cash flow. Infrastructure spending is increasingly supported by debt, leases and external capital. A larger depreciable asset base can weigh on reported earnings, while capital expenditure and financing costs can constrain free cash flow.

The third concerns potential political constraints and power availability. The Democrats’ gains in the midterm elections could intensify scrutiny of electricity costs and the impact on local communities, potentially leading to tighter conditions for new data centres. Alongside grid bottlenecks, water availability and permitting delays, this may slow deployment, increase costs or shift investment towards other locations. The impact will vary considerably depending on project location and readiness. Our base case remains continued expansion of AI infrastructure, supported by a substantial near-term pipeline already under construction, while uncertainty is greater for projects planned further out.

 

Fundamentals continue to support our conviction

Despite these uncertainties, recent results show continued strength in cloud services and AI computing demand. Investment is translating into revenue growth, while semiconductor and equipment suppliers have issued encouraging outlooks for 2027. The current earnings picture therefore remains supportive, even as investors question the longer-term returns.

The divergence between share prices and earnings expectations is particularly relevant. Across several technology and AI segments, prices have fallen while profit forecasts have continued to rise, bringing valuations down. Expectations still need to be delivered, but this combination provides an interesting starting point for investors. Over the six months to 5 November 2026, next-year earnings forecasts rose more strongly for US information technology than for the broader US equity market, with particularly large upgrades in semiconductors and hardware.

 

US earnings forecast revisions: technology versus the broader market

 

Capturing the opportunity across the AI ecosystem

AI’s potential extends well beyond infrastructure. Applications in software development, business automation and scientific research can generate new revenues and improve productivity. More efficient chips and lower computing costs could help broaden adoption and make more applications commercially viable.

At Candriam, we continue to see scope for technology to outperform the broader equity market. We remain positive on semiconductors and technology hardware, while taking a more selective approach to software. Integrated technology platforms also have several ways to monetise AI, through cloud services, existing products and improvements in their own operations.

The investment case will increasingly depend on companies turning adoption into sustainable earnings and cash flow. With demand remaining strong and valuations having adjusted, we believe the sector has scope to outperform the broader equity market, although this depends on companies meeting earnings expectations. In the near term, we will closely monitor the market’s response to Anthropic’s IPO on November 18, and the appointment of a new permanent CEO at Nvidia, following the interim appointment of Colette Kress. Selectivity will be more important than ever.

Key takeaways

  • AI faces greater scrutiny: Investment returns, financing requirements and power availability remain key risks.

  • Earnings offer support: Upward profit revisions underpin our conviction, although trends vary across technology segments.

  • Selectivity remains essential: We favour semiconductors and hardware, with a more selective approach to software.

Look beyond the technology correction

Explore the structural trends shaping AI, automation and digital infrastructure in Candriam’s research on disruptive technologies.

Dernières analyses

  • Envolée de la dette liée à l’IA

    L’essor des infrastructures liées à l’IA pousse certaines des plus grandes entreprises technologiques mondiales à faire davantage appel aux marchés de la dette.
  • Pourquoi nous ne pouvons pas faire l’impasse sur la gouvernance de l’IA

    L’intelligence artificielle (IA) fait désormais partie intégrante de notre quotidien, mais nous commençons tout juste à prendre conscience des risques qui y sont associés. Face à l’incapacité de la réglementation à suivre le rythme, la gouvernance de l’IA constitue désormais un enjeu majeur.
  • Infrastructure numérique prenant en charge l’intelligence artificielle, le <i>cloud computing</i> et le traitement des données

    Rareté, rivalité et nouvelle logique d’allocation d’actifs

    Le passage d’un monde constitué de données « subjectives » à un monde fondé sur des données « objectives » est un changement durable de régime, alimenté par la rivalité géopolitique, la concurrence que représente l’IA et la rareté structurelle des ressources.
  • Tunnel abstrait composé de motifs géométriques bleus répétitifs, convergeant vers un point lumineux au centre.

    Logiciels : L’IA va-t-elle déclencher une SaaSpocalypse ?

    Les agents d’intelligence artificielle vont désormais au-delà de l’assistance pour intervenir dans l’exécution. L’émergence d’agents d’IA comme Claude Cowork a intensifié le débat parmi les investisseurs : une partie de l’écosystème du SaaS pourrait-elle être exposée à un risque de désintermédiation structurelle ?
  • Technologies de rupture

    Le monde de 2030 prend déjà forme : assistants IA intégrés au quotidien, mobilité autonome, usines pilotées par logiciel, jumeaux numériques et systèmes énergétiques intelligents capables d’optimiser la demande en temps réel.
  • L'IA à travers le prisme ESG : Examiner les risques et saisir les opportunités

    La révolution de l’intelligence artificielle (IA) IA est-elle inarrêtable ? À une vitesse fulgurante, l’IA transforme les industries, libère de la valeur économique, génère un impact positif sur l’environnement, contribue à la résolution des défis sociaux - et en outre, offre des opportunités d’investissement convaincantes.

Recherche rapide

Obtenez des informations plus rapidement en un seul click

Recevez des informations directement dans votre boîte e-mail