Nvidia might have to look outside of its artificial intelligence empire to acquire some key data center-related technologies, according to Susquehanna Financial Group. Analyst Christopher Rolland said Nvidia has “two glaring holes” for potential acquisition within optical and analog networking. “As NVIDIA attempts to further solidify its lead in both systems and networking, we believe it may increasingly look to insource or acquire optical and/or analog interconnect technologies, companies, or divisions,” Rolland said in a Thursday note to clients. “As GPU clusters scale linearly, the need for interconnections grows exponentially.” Optical technologies in data centers have a variety of purposes such as supporting network traffic growth, transmitting and receiving optical signals in data centers, and offering cost and power savings while facilitating high-capacity data transmission. Given the growth in cost and importance of these technologies, the analyst expects Nvidia to consider insourcing or acquiring companies that have leading products in linear equalizers, laser drivers, photonic integrated chip technology and digital signal processors, among other areas. Optical and photonic products manufacturer Lumentum Holdings , engineered materials and optoelectronic components maker Coherent and semiconductor producer Marvell Technology are among several names Rolland listed as potential candidates for an Nvidia acquisition. To be sure, the chipmaker has not said it is actively considering aquiring any of these companies. Lumentum shares are up roughly 16.3% this year, while Coherent’s stock price has rallied more than 108% amid the AI excitement, which has led to soaring demand for optical receivers and related technologies. Marvell, meanwhile, is up 18% year to date. Analysts’ average price target, per FactSet, suggests potential downside for Lumentum and Coherent given their recent stock run-ups. The consensus target on Marvell implies shares could gain about 30%, however. Rolland holds a positive rating on Nvidia shares with a $160 price target — higher than the average from analysts polled by FactSet, which implies more than 29% potential upside for the chip darling. He remains bullish on Nvidia’s competitive advantages in semiconductor design, holistic AI systems and its CUDA software toolkit, which he said gives the company a 15-year lead in AI enablement tools, libraries, and operating software.