Gary Smith, chief executive of optical-networking company Ciena, draws on his experience navigating the telecommunications crash of the early 2000s to assess today’s AI-driven technology surge, while cautioning against viewing the current moment as a repeat of past boom-and-bust cycles.
Smith, who has led Ciena since May 2001, recalls the turbulence that followed the dot-com bust, highlighting an industry plagued by an oversupply of fiber-optic cable, sharp price declines, a frozen venture capital market, extensive bankruptcies, and widespread job losses. Ciena itself endured dramatic fluctuations during this period, with a market capitalization that soared above $40 billion before plunging more than 90%. The company survived and has since recovered, currently boasting a market value near $45.75 billion and ranking fifth in a recent board effectiveness assessment.
Despite superficial similarities between the telecom buildout and the current proliferation of artificial intelligence applications, Smith emphasizes important distinctions that set the AI surge apart. He characterizes the AI era as “a long ramp of compounding waves of application and demand,” rather than the faster boom-and-bust cycles typical of previous technology phases.
According to Smith, the initial wave involves connecting multiple data centers, a process still underway. The second wave, he says, is just starting and centers on inference computing—where businesses and individuals apply pre-trained AI models to address real-world challenges. Later waves are expected to involve AI agents that autonomously perform tasks of increasing complexity, followed by advances in robotics or “physical AI.” Smith asserts that these developments place growing demands on network infrastructure, which he believes has been underestimated in scale.
To meet this demand, Ciena is advancing cutting-edge technology, including proprietary 3-nanometer WaveLogic chips capable of transmitting data at 1.6 terabits per second over a single fiber strand, with 2-nanometer versions in development. Smaller transistor sizes allow more to fit on a chip, improving performance and efficiency.
Industry analysts note key structural differences between the early 2000s telecom market and today’s AI-focused environment. Heath Terry, global head of technology and communications research at Citi, points out that about 94% of fiber installed during the telecom boom remained unused, whereas “there are no dark GPUs” fueling AI workloads. Smith also highlights the stronger customer base today, including profitable cloud computing leaders, contributing to sustained demand that Ciena has struggled to fully satisfy.
Looking ahead, Ciena projects a compound annual growth rate of 30% from fiscal 2026 through 2029 and anticipates ending fiscal 2026 with roughly $10 billion in order backlog, indicating continued momentum driven by expanding AI-related infrastructure needs.